Compare / Qwen3.6-Plus vs Qwen3.7-Plus

Qwen3.6-Plus vs Qwen3.7-Plus

Which is cheaper, which is more capable, and which does better at each kind of work, from 17 results on 2 sources that measured both models.

AlibabaQwen3.6-Plus Released 31 Mar 2026 · $0.33 in · $1.95 out
AlibabaQwen3.7-Plus Released 2 Jun 2026 · $0.32 in · $1.28 out
Qwen3.6-Plus wins 53 too close to callQwen3.7-Plus wins 9

Of the 17 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.

Results as of 19 November 2026 (catalogue release 2026-11-19-3608133e8331).

Qwen3.6-Plus or Qwen3.7-Plus? The short answer

Qwen3.7-Plus is the newer version; they score about the same, so pick Qwen3.7-Plus, which costs less (23% less per token).

  • Qwen3.7-Plus is cheaper: $0.56 against $0.73 per million tokens, 23% less.
  • Qwen3.6-Plus scores higher on the Artificial Analysis Intelligence Index: 27.0 against 25.2.
  • Qwen3.7-Plus is newer: released 2 Jun 2026, against 31 Mar 2026 for Qwen3.6-Plus.
  • On third-party benchmarks, on the results both models have: Qwen3.6-Plus does better for agents and tool use; Qwen3.7-Plus for reasoning and knowledge, writing and sticking to the facts.

Head to head

27.0
Intelligence Index
25.2
$0.73
Price per million tokens, blendedlower is better
$0.56
1,000,000
Context window, tokens
1,000,000

Every fact side by side

Qwen3.6-PlusQwen3.7-Plus
Price per million tokens$0.33 in · $1.95 out$0.32 in · $1.28 out
Blended (3 in : 1 out)$0.73$0.56
Cheapest host, blended–$0.56 (Alibaba)
Intelligence Index27.025.2
Spring Prompt overall––
Output speed, tokens a second–56
Context window1,000,000 tokens1,000,000 tokens
Released31 Mar 20262 Jun 2026 (newer)
WeightsClosedClosed
Developer based inChinaChina

Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.

Which is better for what

Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.

Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 68 of 159 77 of 159
Reasoning and knowledgeHard questions across science, maths and general knowledge 73 of 187 64 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 55 of 169 48 of 169
Sticking to the factsSummarising without inventing things, and factual answers 77 of 139 52 of 139

Every shared result

Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.

MeasureQwen3.6-PlusQwen3.7-Plus
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 76 · 59 of 387 measured 1,443 1,456
Overall · Text factualityArena rating, higher is better 84 · 52 of 163 measured 1,438 1,453
Business, management and finance · TextArena rating, higher is better 75 · 77 of 380 measured 1,446 1,445
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 66 · 79 of 249 measured 27.0 25.2
GPQA Diamond% of questions, higher is better 59 · 42 of 232 measured 88.2% 90.0%
Humanity's Last Exam% of questions, higher is better 92 · 63 of 250 measured 27.8% 35.6%
Show all 17 results (11 more)
MeasureQwen3.6-PlusQwen3.7-Plus
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 4 results
Creative writing · TextArena rating, higher is better 82 · 60 of 385 measured 1,407 1,432
Expert prompts · TextArena rating, higher is better 68 · 56 of 338 measured 1,476 1,485
Instruction following · TextArena rating, higher is better 74 · 58 of 387 measured 1,435 1,447
Writing, literature and language · TextArena rating, higher is better 76 · 61 of 386 measured 1,423 1,437
Artificial Analysis reported by Artificial Analysis · 7 results
Terminal-Bench 2.1% of tasks, higher is better 48 · 49 of 131 measured 61.4% 61.0%
τ-bench banking% of tasks, higher is better 48 · 52 of 121 measured 20.8% 17.5%
Artificial Analysis Coding Indexindex score, higher is better 50 · 48 of 130 measured 54.5 55.9
IFBench% of instructions, higher is better 29 · 9 of 195 measured 75.2% 78.0%
Long-context reasoning (AA-LCR)% of questions, higher is better 59 · 100 of 241 measured 78.3% 73.0%
Terminal-Bench Hard% of tasks, higher is better 23 · 15 of 190 measured 43.9% 47.0%
τ²-bench telecom% of tasks, higher is better 7 · 32 of 193 measured 97.7% 93.0%

Quick answers

Is Qwen3.6-Plus better than Qwen3.7-Plus?

Qwen3.7-Plus is the newer version; they score about the same, so pick Qwen3.7-Plus, which costs less (23% less per token).

Which is cheaper, Qwen3.6-Plus or Qwen3.7-Plus?

Qwen3.7-Plus is cheaper: $0.56 against $0.73 per million tokens, 23% less.

Which is better for agents and tool use, Qwen3.6-Plus or Qwen3.7-Plus?

Qwen3.6-Plus ranks 68 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 77 for Qwen3.7-Plus.

Which is better for reasoning and knowledge, Qwen3.6-Plus or Qwen3.7-Plus?

Qwen3.7-Plus ranks 64 of 187 models on sale for hard questions across science, maths and general knowledge, against 73 for Qwen3.6-Plus.

Which is better for writing, Qwen3.6-Plus or Qwen3.7-Plus?

Qwen3.7-Plus ranks 48 of 169 models on sale for what people prefer in blind comparisons, and judged writing quality, against 55 for Qwen3.6-Plus.

Which is better for sticking to the facts, Qwen3.6-Plus or Qwen3.7-Plus?

Qwen3.7-Plus ranks 52 of 139 models on sale for summarising without inventing things, and factual answers, against 77 for Qwen3.6-Plus.

Run Qwen3.6-Plus and Qwen3.7-Plus on your own prompt.

Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.